Steve Baker (steve++at++mred.bgm.link.com)
Wed, 11 Dec 96 08:32:40 -0500
We have a 3 pipe system with APP_CULL_DRAW, every process on a 200MHz CPU
of it's own. Every process is running comfortably less than ~24msec
per frame and the system is correctly GenLocked (as far as we can tell).
Yet despite all of this, it refuses to run at 30Hz.
Timings are:
DRAW: 22, 24, 22ms
CULL: 7, 10, 7ms
APP : 10ms (end of pfFrame to pfSync)
<1ms (end of pfSync to pfFrame)
Unfortunately, we are have very tiny channel sizes and using
Performer stats is hard. However, careful timing of my own clearly
reveals that the DRAW processes on pipes 0 and 1 are not starting
up until well into the frame. This results in them failing to complete
within 33msec. Pipe 2's DRAW, and all three CULL processes are
starting at exactly the correct time (very shortly after pfFrame
returns in APP) but the DRAW processes for pipes 0 and 1 are not
starting until sometime AFTER their respective cull processes have
completed since the times for CULL and DRAW taken together exceed
33ms, we break frames all the time unless we move to a very
simple area of the database such that CULL+DRAW can get under 33ms.
We are running with the pfPhase set to LIMIT - FREE_RUN
seems to do the same thing - and the system locks up completely
in LOCK to the point that we have to do a stopgfx/startgfx to
get the system to come back.
I don't think this is a new problem, we have seen it before,
but managed to 'get away with it' because CULL times were a
lot smaller.
HELP!!
Steve Baker 817-619-1361 (Vox-Lab)
Hughes Training Inc. 817-619-8776 (Vox-Office/Vox-Mail)
2200 Arlington Downs Road 817-619-4028 (Fax)
Arlington, Texas. TX 76005-6171 Steve++at++MrEd.bgm.link.com (eMail)
http://www.hti.com (external) http://MrEd.bgm.link.com/staff/steve (intranet)
http://web2.airmail.net/sjbaker1 (external)
"You can't destroy the Earth - that's where I keep all my stuff!" - The Tick.
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
Submissions: info-performer++at++sgi.com
Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:54:08 PDT